COMPLIANCE
Europe ’ s quiet AFC revolution
AI has been the primary buzzword in recent years when it comes to fighting financial crime . In 2021 , the Financial Action Task Force ( FATF ) published the report titled “ Opportunities and Challenges of New Technologies for AML / CFT ” in which it describes the advanced computational techniques of AI that include pattern recognition , predictions and decisions . 6
The report , however , fails to mention quantum computing , but several European-led projects are driving a quiet revolution that could transform the tools used in AFC .
In 2022 , a research project conducted by members of the European Organization for Nuclear Research ( CERN ), IBM Zurich and the German sensor component manufacturing firm , Iris GMBH , found that quantum support-vector machine algorithms ( QSVM ) can surpass AI when it comes to fraud detection . 7
QSVM algorithms are a type of machine learning algorithm that use quantum mechanics to classify data into different categories . Classical SVMs take data that is jumbled ( i . e ., pictures of cats and dogs , apples and oranges or , in our case , data related to suspicious and nonsuspicious transactions ) and draw a line , known as a hyperplane , between the groups so it can separate the data into two or more distinct categories . 8 By adding properties of quantum mechanics , the algorithm can handle larger and more complex data sets , and perform calculations faster than a classical SVM . 9
The QSVM , created by the research project and run on an IBM quantum computer , sifted through 2.4 million credit card transactions to detect fraudulent activity . The researchers discovered that the quantum classifiers were superior at identifying patterns of data compared to classical SVMs . They also found that the QSVM algorithm struck a better balance between false positives and false negatives , the bane of most AFC analysts . 10
Similarly , the Spanish start-up Multiverse Computing created a quantum algorithm that combed through 300,000 payments and identified 500 fraudulent transactions . The result was found to be 2 % more accurate than traditional algorithms . 11 This may not seem much , but for a large financial institution ( FI ), this means saving millions of euros in damages .
Early adoption
FIs across Europe have already begun experimenting with the use of quantum computers for AFC .
In December of 2022 , Dutch multinational FI Rabobank partnered with the U . S . startup Quantum Computing Inc . to add to its existing arsenal of fraud detection techniques a Dirac 1 entropy quantum computing ( EQC ) system . 12
Most quantum computers are prone to “ noise ,” or environmental changes — such as electromagnetic fields or heat — that can disrupt qubits . An EQC system engineers the environment for increased error-free quantum processing . 13
The Dutch colossus is not alone . In 2019 , Spanish CaixaBank put together a team of IT technicians , mathematicians and others to explore quantum technologies to enhance an FI ’ s capabilities , including fraud detection . It is currently working with American startup D-Wave on quantum applications . 14
The EU ’ s race for quantum advantage
FIs in Europe and across the world are dependent on multinational technology corporations , such as IBM or Google , for the use of quantum
56 acamstoday . org